Abstract
Contingent capture cueing paradigms have long shown that salient visual stimuli – both abrupt onsets and color singleton cues – fail to reliably capture attention if they do not resemble the search target. Recent evidence indicates that abrupt-onset cues may capture attention in this paradigm in easy, but not difficult, search displays. Regardless of why this is, we hypothesized that if a cue reliably captures attention, people should be able to pick up on any regularities between it and the target location. That is, it should be possible to expose any latent capture generated by the cue through statistical learning. To test this, we ran two versions of the contingent capture paradigm with easy search displays. In both versions, the target was defined by its unique color in an array of distractors and was preceded by a cue, which matched the target feature in one condition (match cue). The mismatch cue in Experiment 1 was a color singleton and in Experiment 2 it was an abrupt onset. Unbeknownst to participants, in both experiments, the target predictably appeared at the same location as the mismatch cue (81.5%) whereas the match cue did not predict the target location (25%). Replicating typical findings, capture was robust and stable over time for match cues. Mismatch color cues consistently failed to produce capture throughout the experiment. Importantly, mismatch abrupt-onset cues did produce capture after the first block of trials (i.e., after statistical learning). The dissociation exposes latent capture by abrupt-onset cues in easy search. Together, the findings suggest that attentional control sets are not so powerful that all information is filtered out while also showing that statistical learning is not so powerful that it undermines all top-down control.
Acknowledgement: Natural Sciences and Engineering Research Council